Translocation of FGF1 and FGF2 factors to the cytosol and nucleus

Supported by grant from Norway through the Norwegian Financial Mechanism

Engineering of the aFGF - Fibroblast Growth Factor

(in colaborations with Prof. Antoni Wiedlocha)

Human acidic fibroblast growth factor (FGF-1) belongs to the large family of FGF growth factors. It is a powerful mitogen, involved in stimulation of DNA synthesis, inducing the proliferation of a wide variety of cell types, and playing an important role in morphogenesis, angiogenesis and wound healing. The well-established mode of action of the growth factor consists of receptor binding and activation of the tyrosine kinase domain in the cytoplasmic part of the receptors. There are reports indicating that exogenous FGF-1 is capable of reaching the cytosol and the nucleus of target cells. Accordingly, signaling occurs partly through cell surface receptors and partly by transport of the growth factor into the cell.

Despite years of intensive research, the mechanism of translocation and intracellular interactions of FGF-1 has remained unclear. FGF-1 is inherently poorly stable protein and the low stability of the wild type of FGF-1 may be related to regulatory mechanisms, especially cell membrane translocation. FGF-1 has a b-trefoil structure with a centrally located hydrophobic internal cavity.

Using homology modeling, sequence and structure comparisons of 150 members of the FGF family, we design and construct single and multiple mutants of FGF-1, determine their thermodynamic stability, and biological activities. The mutations are introduced by site-directed mutagenesis, expressed using plasmid/Bl21(DE3)pLysS host expression system and purified by affinity chromatography.

The first step of the study is to obtain FGF-1 single mutants of increased stability by optimizing the core packing by central cavity rearrangement, creation of new hydrogen bonds/salt bridges and by introducing negatively charged or neutral side chains at the surface clusters of positively charged amino acids. The effects of the mutations are determined by thermal and solvent denaturation. Further increase of stability of FGF-1 should be possible by combination of most stabilizing single mutations and production of double or multiple mutants.

For all the mutants we perform tests of biological activities such as: stimulation of DNA synthesis, MAP kinase activation, receptor binding, in vivo phosphorylation. In order to clinically apply FGF-1 variant all these tests should be positive, at least at the level of the wild type growth factor. We would like to explain of the role of kinases in signal transduction mechanism and also to give a new perspective to FGF-1 mode of action inside the cell. Final step in our research will be crystallization and x-ray structure determination of several FGF-1 mutants showing the most interesting biological and stability properties.

Generation of proteinase inhibitors from camel antibody library

New inhibitor for coagulation factor Xa and a-thrombin derived from VHH domain of camel heavy-chain antibodies.

Old world camels (Camelus bactrianus and Camelus dromediarius) and new world camelids (Lama pacos, Lama glama and Lama vicugna) have unique humoral immune response. They possess the conventional heterotetrameric antibodies and also have heavy-chain antibodies in their serum (CHAB). Functional molecule of CHAB is composed of two heavy chains only. In each chain three domains are present (VHH, CH2, CH3) with a hinge region between VHH and CH2 domain. There are two subclasses of CHAB:
IgG2 - heavy-chain antibodies with long hinge region (92 kDa);
IgG3 - heavy-chain antibodies with short hinge region (90kDa).

Heavy-chain antibodies posses half of antigen binding activity residues compared with 'classical' antibodies. Nevertheless, in the lack of VL domains the CHABs recognise antigen with the high affinity. There are number of adaptations inside a single variable heavy chain domain (VHH) to cope with the absence of the variable N-terminal light chain domain (VL) from the antigen-binding site. One of these adaptations is usually extremely long and highly variable third complementarity determining region (CDR3). Long loop of the CDR3 can penetrate and bind to the antigen in its cavities, clefts or grooves (for example region of enzyme active site).
There were few reports where camel heavy-chain antibodies reached and interacted with active sites of enzymatic-antigen. This is why they can be applicable in designing new enzyme inhibitors.

The idea of the project is to generate novel, small protein inhibitors against: human coagulation factor Xa, a-thrombin, gingipain, caspase 3 and caspase 8; derived from heavy chain camel antibody (CHAb). We proceeded with three independent animal immunizations:
1. Camelus dromedarius 49 - was immunized by hFXa and a-thrombin;
2. Lama guanicoe - was immunized by hFXa and a-thrombin, gingipain, caspase 3 and caspase 8;
3. Camelus dromedarius 52 - was immunized by gingipain, caspase 3 and caspase 8;

In each case B-lymphocyte cells were extracted from peripheral blood. B-cell mRNA was purified and reversibly transcribed to cDNA, which encoded a repertoire of single variable heavy chain domains (VHH) of about ~107 variants. We used phagemid system pHen4 to construct three cDNA libraries:
1. One library for Camelus dromedarius 49 (which contained ~4x107 independent transformants);
2. Two libraries for Lama guanicoe:
- The first in original pHen4 system (which contained ~1x108 independent transformants);
- The second in modificated pHen4 system, where the amber STOP codon between VHH and pIII was eliminated (which contained ~1.8x107 independent transformants);
3. Library for Camelus dromedarius 52 is under construction;

We plan to:
- Perform several selections of the libraries using phage display technique;
- Identify selected clones;
- Express positive binders;
- Characterize interaction between VHH and target proteins;

C-terminal phage display

PDZ (PSD-95/Discs-large/ZO-1) domains are protein-protein interaction modules, commonly found in signaling proteins. PDZ domains contain about 90 residues and fold into a structure which consists of six beta-strands and two alpha-helices specialized in binding of C-termini of partner proteins, most often transmembrane receptors and channel proteins, and/or other PDZ domains. Despite numerous crystal and solution structures determined for various PDZ domains both in peptide-free and peptide-bound states, molecular mechanism of their specificity remains unclear.

Phage display is a technique widely used in investigating protein binding specificity. Unmodified filamentous phages allow display of peptides attached to the amino-termini of capsid proteins. Our approach is directed specifically towards studying PDZ domains and thus peptide library has to be displayed C-terminally. We are using the C-terminal phage displayed library to study binding specificities of PDZ domains from several mammalian proteins (dishevelled, LARG, PDZ-RhoGEF, MUPP, syntenin). Peptide variants binding to those PDZ domains are further investigated using biophysical methods.

Muskelin

Muskelin is a multidomain intracellular protein discovered in a mouse muscle cell line (Fig.1). It is supposed that is involved in cytoskeletal organization and neuronal migration. It was initially characterized as having functional involvement in cell responses to thrombospondin-1. This process is integrated into the regulation of cell-adhesive behaviour and cytoskeletal organization.

Bioinformatic analysis showed that N-terminal region of muskelin contains F58C domain (also known as discoidin domain) (Fig.2). This domain, identified in a number of extracellular and intracellular proteins and in clotting factors V and VIII, mediates phospholipid binding. Discoidin domain plays also important function in a protein-protein interaction and self-association of proteins. The C-terminal part of muskelin is predicted to have six repeated kelch motifs and undefined short C-terminal region. Each kelch motif forms a four-stranded antiparallel beta-sheet which forms form a blade of beta-propeller structure (Fig.3). This module contains protein-protein contact interface.

The intracellular muskelin shows structural similarity to extracellular fungal galactose oxidase. It is not clear what structural elements decide on completely different function of both proteins. The central region of muskelin is predicted to have alpha-helical secondary structure and contains two sequence motifs, the LisH motif and the CTLH motif. These motifs distinguish structure of muskelin from structure of galactose oxidase and other kelch-repeat proteins.

Our aim is to determine the complete structure of muskelin and identify "hot spots" of protein-protein interactions. We would like to establish the role of each domain or/and tandem of domains. Further study will be concentrated on cellular function of muskelin and its domains and biological consequence of the point mutations.

Elfin

Elfin is a protein that associates with actin filaments and stress fibers through interaction with alpha-actinin. Actin stress fibers form a dynamic organelle that is important in processes involving cell shape, such as cell migration, cell polarity and cytokinesis. Upon the cell activation elfin and alpha-actinin translocate as a complex from the cytosol to the F-actin-rich cytoskeleton.

Elfin consists of protein interaction motifs that are found in various proteins associated with the cytoskeleton. It has been suggested that elfin acts as an adapter between kinases and the cytoskeleton. Elfin consists of three domains. The N-terminal PDZ domain is formed by six beta-strands and two alpha-helices. It is found in various sub-membranous and cytoplasmic proteins. PDZ motif binds the carboxyl-terminal sequences of proteins or internal peptide sequences. PDZ-containing proteins are known to function in the clustering of ion channels and receptors. In elfin it is thought to bind to spectrin-like repeats of alpha-actinin 1 and alpha-actinin 4.

The second domain of elfin includes a short internal sequence (called ZM domain). Its plausible function is to direct intracellular proteins to multiprotein complexes. In homologous proteins it is crucial for their proper localization in the cell. Further studies are underway to elucidate whether ZM domain represents a conserved interaction site or it is mainly a structural motif.

The third C-terminal LIM domain is found in many key regulators of developmental pathways including homeodomain-containing transcription factors, cytoskeletal-component proteins and other signalling molecules, that are known to be involved in cell-fate determination, growth regulation, oncogenesis and cytoskeletal reorganization. It contains a conserved cysteine-rich domain that binds two zinc ions. LIM does not bind DNA, rather it seems to act as an interface for protein-protein interaction. In elfin it interacts with the C terminal EF hands of alpha-actinin 2.